Research On Cluster-Based 3D Massive MIMO Channel Measurements And Modeling

With the rapid updates of wireless communication systems,the demand for in-depth understanding of wireless channel,i.e.,the medium of wireless communications,is also increasing.Throughout the history of human commercial mobile communications,the wireless air interface technology has evolved from analog communication of the 1st generation mobile communication(1G),to the Time Division Multiple Access(TDMA)of the 2nd generation mobile communication(2G)and the Code Division Multiple Access(CDMA)of the 3rd generation mobile communication(3G),then to the Orthogonal Frequency Division Multiplexing(OFDM)and Multiple Input Multiple Output(MIMO)of the Long Term Evolution(LTE).The wireless resources in time,frequency,code,and space domains have been greatly expanded and utilized.In the 5th Generation Mobile Communication(5G),the popularization of a large number of mobile intelligent terminals,large-scale applications of Internet of Things devices,and the emergence of new services in the vertical industry will lead to explosive growth of the mobile data traffic,and further put forward severe requirements for the new wireless air interface technology.At present,massive Multi-Input Multi-Output(massive MIMO),also known as large-scale MIMO,has been selected as one of the key technologies of 5G.Theoretical research shows that massive MIMO can greatly improve spectrum efficiency,increase system transmission rate and throughput,reduce energy loss,and improve communication reliability.When combined with beamforming technology,it can fully make use of the spatial resources of the wireless channels,and produce significant array,interference suppression,and space division multiplexing gains.Understanding wireless channels of massive MIMO is the precondition for the design,analysis,evaluation and application of massive MIMO wireless communication systems.The mastery of the radio propagation characteristics of massive MIMO is the physical basis for the research of massive MIMO wireless communications.A reasonable and accurate channel model which can truly reflect the characteristics of massive MIMO channel is the premise of technical verification and performance evaluation of massive MIMO prototype.Channel modeling for massive MIMO communications has aroused great interest in academia and industry.However,deep research on channel characteristics and channel models for massive MIMO is still limited so far.In order to promote massive MIMO from theoretical research to practical application,there is an urgent need for more channel characterization and channel modeling for massive MIMO communications in different hot scenarios.For the motivations mentioned above,this dissertation investigates the channel characteristics and studies the channel modeling method of massive MIMO channels,based on a series of channel measurement campaigns and simulations for massive MIMO channels.The main contributions and novelties of the dissertation are as follows.(1)An automatic channel sounder system for massive MIMO channels is developed.The sounder system adopts virtual array technology and can support different three-dimensional(3D)array topologies,such as linear array,rectangular array,and cylindrical array,etc.It supports a wide frequency range(3.33 GHz-26 GHz).Using the developed automatic channel measurement system,massive MIMO channel measurement campaigns in different typical hotspot indoor scenarios are carried out.These measurements laid a solid foundation for massive MIMO channel characterization and modeling.(2)A novel clustering framework is proposed to cluster the multipaths over the large-scale array.In the hybrid approach,we apply the Space-Alternating Generalized Expectation maximization(SAGE)algorithm to estimate the multipath components(MPCs),and use the Multipath Component Distance(MCD)-based tracking algorithm and the KPowerMeans algorithm for MPCs tracking and clustering.Moreover,the concepts of common cluster and non-common cluster are proposed.A cluster partition algorithm is further proposed to adjudge the clusters association over the array,and output the final clustering results.(3)Based on the clustering results,comprehensive channel characterizations are carried out in delay,frequency,angular,and array domains.The 3D propagation mechanisms of the channels are investigated.Specifically,the distribution of MPCs in the elevation domain is investigated,and the channel spatial non-stationarity in both of the horizontal and vertical array dimensions is characterized.Moreover,the spherical wavefront propagation model of the channel is derived based on solid geometry.(4)Dense multipath component(DMC)modeling for massive MIMO channels is performed based on measurement campaigns in three different indoor hotspot scenarnos.The parameters for DMC are extracted by applying a RiMAX-based estimator.The distance-dependence and the variance behaviors of the DMC parameters across the large-scale array are analyzed and modeled.DMC implementation and validations are presented.Validation results show that the propose DMC model can significantly improve the accuracy of the massive MIMO channel modeling.(5)A 3D cluster-based channel model for massive MIMO communications based on our realistic channel measurements is proposed.Several global parameters,e.g.LOS and shadowing components,overall angle distribution,global angular spread,and a set of inter-and intra-cluster parameters are provided.In addition,detailed the model implementation and validation are presented.The entropy capacity and Demmel condition number are used as the validation metrics,showing good match between the measured and simulated channels.